Groups Clash Over Discrimination Standard in Automated Valuation Rule

Learn why credit union trade groups are at odds with consumer groups over a proposed rule governing Automated Valuation Models in the appraisal industry.

David Baumann


Aug 23



View all posts by 

David Baumann

Articles Posted by

David Baumann

A squiggly pink arrow pointing downward and to the right.

Credit union trades, consumer groups at odds over proposal.

Humans may produce property appraisals that are racially or ethnically discriminatory, but can machines?

And if they can, who should be held responsible for the bias?

Those seem to be the central questions dividing commenters on the federal financial regulators’ interagency proposed rule governing Automated Valuation Models (AVMs).

Credit union trade groups contend that their member credit unions should not be responsible for ensuring models do not discriminate.

But consumer groups argue a discrimination standard is essential to ensuring that property owners receive an accurate evaluation.

Backstory and Context

The federal financial regulators—including the NCUA and the CFPB—issued the proposed rule in June, with comments due Monday.

“While advances in AVM technology and data availability have the potential to contribute to lower costs and shorter turnaround times in the performance of property valuations, it is important that institutions using such tools take appropriate steps to ensure the credibility and integrity of the valuations produced by AVMs,” the agencies said, in issuing the rule.

Under the proposal, mortgage originators and secondary market issuers would be required to adopt measures to ensure that AVMs adhere to quality control standards. However, the rule would not establish specific requirements for how financial institutions set quality control standards based on the size of the institution and the complexity of the transactions that use AVMs.

Several groups applauded the agencies for providing credit unions and banks with that flexibility.

Inside the Comments


CUNA supports the agencies’ principles-based approach, wrote David Park, the trade group’s senior director of advocacy and counsel, in commenting on the proposal.

“Credit unions favor that approach because it ensures they will be able to tailor their quality control standards to their unique circumstances and account for their size, complexity, and risk profile,” his letter reads.

However, Park said that credit unions—particularly small ones—are concerned that the regulatory burden of requiring a non-discrimination quality control standard would preclude them from being able to use AVMs.

He added that credit unions might not have sufficiently large data sets to perform statistically significant testing with their AVM results.

“The proposed non-discrimination quality control factor injects a level of uncertainty that could lead some credit unions to decide that the most prudent course of action going forward is to stop using AVMs in the mortgage origination context,” Park stated.


NAFCU also supports the concept of a principles-based approach to AVM quality controls, according to James Akin, the trade group’s senior regulatory affairs counsel.

However, he said AVM providers—and not credit unions—should have the responsibility to ensure that they comply with nondiscrimination laws and stakeholders should establish a Standards Setting Organization to develop specifications for reliable AVMs.

Akin noted that NAFCU member credit unions have reported that they have seen an increase in the price of appraisals and further that there is a lack of standardization among appraisers.

“The ability to use AVMs as well as desktop appraisals or other alternatives and appraisal waivers in rural and underserved areas will decrease the strain placed on NAFCU’s member credit unions and their member-borrowers when financing the purchase of a home through the mortgage loan process,” he wrote.

However, he added that ensuring compliance with anti-discrimination laws can pose a significant challenge for credit unions, and that rather than including an anti-discrimination requirement in the rule, agencies should instead provide clear guidance on relevant red flags in AVMs. This, according to Akin, would put AVM developers on notice that credit unions have the tools needed to evaluate whether AVMs are discriminatory.

National Fair Housing Alliance and National Consumer Law Center

These two groups submitted a joint comment letter, stating that it is “critically important” that discrimination be included in the final rule.

“AVMs run a high risk of perpetuating discrimination if they are not adequately examined and tested: there is no question that discriminatory mis-valuations are a historical and present phenomenon,” they wrote.

SBA Office of Advocacy

The SBA Office of Advocacy acts independently from the agency in evaluating proposed rules.

Attorneys from the office told the agencies that small entities do not control AVM data and therefore do not have the ability to control the quality of that data. In addition, small entities do not have the bargaining power of large companies, so they may lack the influence to make demands of an AVM vendor.

“As such, it is unreasonable to hold them responsible for the actions of AVMs,” they concluded. “Advocacy recommends that small entities be exempt from the rulemaking.”

Real Estate Companies

The president and CEO of Veros Real Estate Solutions commended the agencies for issuing a proposal that is non-prescriptive and flexible.

“The approach allows for important guidelines of reasonable quality control standards to be put in place,” wrote Darius Bozorgi, in commenting on the proposal.

He said his company has studied racial bias in its use of its AVM and found no empirical evidence that its valuations are influenced by race or other variables related to disadvantaged communities.

However, officials from HouseCanary, a national brokerage firm, said that inclusion of a discrimination standard is “one of the notable achievements” in the proposed rulemaking.

“At HouseCanary, we wholeheartedly support the addition of this factor and have consistently conducted rigorous testing to identify and address any potential discrimination in our model outputs,” they wrote.

Industry News

No items found.